Centrifugal Compressor Curves Correction

[essenwood]

Hi all,
Hope someone can help me...

I extracted this from a centrifugal compressor book while trying to find out how to do corrections to performance maps of centrifugal compressors.

"The performance map for a centrifugal compressor typically supplied by the vendor would be either polytropic head, pressure ratio, discharge pressure on the y axis. On the x axis would be the actual inlet flow such as acfm or icfm where a = actual ; i = inlet. This volume or capacity is based on a gas with a particular molecular weight, specific heat ratio, and compressibility factor at a pressure and temperature corresponding to the gas condition in the suction line to the compressor. If any of these parameters changed, the performance map is no longer excatly valid.
If the chnages are minor, the map can still be used with a fair amount of accuracy by making small adjustments."

My question is what is the adjustments or calculations involved to correct for the deviation form the inlet design conditions?....appreciate if some formulas and numbers can be thrown into the explanation.

Thanks
Essenwood.

 

[pmover]

essenwood,
the statement: "If any of these parameters changed, the performance map is no longer excatly valid.
If the chnages are minor, the map can still be used with a fair amount of accuracy by making small adjustments." " needs to be quantified with in so far as changing inlet conditions. inlet flow to a centrifugal compressor is critical to avoid operating the compressor in an unsafe condition (i.e. surge - can be catastrophic or stonewall - depending upon compressor, it may/may not be detrimental to compressor; however, certainly not efficient at all).
however, if inlet conditions are known and a compressor map is available as you described, one can reasonably predict the outlet conditions. please refer to earlier post "Compressor head curve" thread407-60637.
heck, you posted that one as well! let's see, how can i expand on your situation.

if inlet conditions are changed beyond what the mfg recommends for previous compressor performance curves, then a prudent engineer should obtain a new set of curves based on revised inlet conditions from the oem (original equipment mfg)!

it is difficult for the operating engineer to apply any corrections to compressor performance without knowing or having oem data. at one time, elliott and delaval had good technical user guides for users to utilize in estimating compressor sizing and performance, but whether or not they still publish and issue them, i do not know.

perhaps you could be a little more specific about what your concern is or are questioning!

determining performance of centrifugal units is a relatively straightforward process.
please advise and good luck!
-pmover

 

[essenwood]

Hi Pmover,
Sorry if I ticked you off...it's not that I am questioning your wisdom.
I tried to find a way to contact you via email; but I don't think it is posted anywhere in the forum. I also posted a reply in a last thread but unfortunately you were not around.


The manufacturer told me that for minor changes in the inlet conditions; I don't have to worry about it as the curves would be ok (as expected..they don't to be bothered with new curves for every single set of conditions).

From the last thread that you contributed, I kindda get it , but I'm not sure...if you could show me some examples of how it can be done...

We were doing a surge test with the vendor...from that test a point would be identified as a surge point (and a line) for the compressor. That point would be located on a map (that mapp would be the manufacturer supplied one). That is all ok with me until my curiosity wasted me valuable time:

The inlet conditions would definetely change, seasonal, tower operating conditions, feedstocks..my curve would need to be corrected...would my surge point/line identified in the test still be valid...


Thanks Pmover...help me out of this...I have wated enough time thinking about this.


Essenwood

 

[pmover]

ok!

absolutely not! you did not disappoint/upset me. just trying to understand what is needed.

it is good to know that you are learning and i've done several surge tests at other than design conditions, so no concern about using oem performance curves.

the performance curves to utilize are the head (polytropic or isentropic) vs inlet flow (clarify units!-cfm, acfm, icfm).

obtain inlet and outlet conditions (i.e. pressure, temperature, and gas composition), obtain flow from fmd (flow measurement device) and correct to inlet conditions at compressor (perhaps this is where u may be confused! as p & t at fmd may be different than at compressor!), and obtain compressor rpm.
compute head per recommended mfg method (either polytropic or isentropic)
plot point on performance curve.
the plotted point should nearly match the rpm of the compressor at the inlet flow conditions.
good luck!
-pmover

 

[pmover]

after further correspondence, i understand the matter in which the compressor may operate at different moleweight conditions than what the original compressor mfg predicted performance curves were established. in this case, it is best to consult with the oem and obtain a set of predicted performance curves for new moleweight/process conditions and then confirm by conducting a test. the alternative is to test the compressor as it is and create your own set of performance curves and then determine/verify difference between original and current curves is a change in moleweight.
either choice will provide useful information to the operator.
good luck!
-pmover

 

[KiwiBob]

There are many locations to find this sort of detailed information (like the various centrifugal compressor vendors), but the one fundamental piece of advice that I would give you is that an impeller has one characteristic curve for a given speed, when actual inlet volume (in say m3/h)is plotted against head (in say kJ/kg or m). I would phrase it as "all the impeller knows is inlet volume flow and head rise". If you work from that basis, variations in from a given design point can be worked through.